Advancing CO₂ capture with the CESAR1 solvent: experimental characterization and process modeling

Climate change, primarily driven by high atmospheric CO₂ concentration, is one of the most pressing environmental challenges of this century. Carbon capture technologies are expected to play a significant role in the decarbonization of hard-to-abate industrial sectors. Among these technologies, chemical absorption using aqueous amine systems is the most mature method for capturing CO₂ from a broad range of industrial flue gases. Many amine systems have been studied over the years, with ethanolamine (MEA) being among the most researched ones and traditionally considered the benchmark for evaluating new solvents. A blend of 3 M 2-amino-2-methyl-1-propanol (AMP) and 1.5 M piperazine (PZ), named CESAR1, has been recently proposed as a new benchmark due to its higher energetic efficiency and chemical stability compared to MEA. However, CESAR1 was investigated less extensively than MEA. Therefore, this thesis focuses on improving the fundamental understanding of CESAR1 performance for amine-based CO₂ capture. A literature review of the properties of CESAR1 and its individual components highlighted experimental gaps both at laboratory and pilot-plant scales. This thesis addresses experimental laboratory data gaps that hinder the development of rigorous process modeling for CESAR1-based CO₂ capture design and operation. However, the work is part of a bigger project, AURORA, in which other challenges, such as amine degradation, solvent emissions, and pilot testing, are also addressed. Density and viscosity were measured for aqueous AMP/PZ mixtures, including the CESAR1 formulation, over a broad range of temperatures, amine concentrations, and CO₂ loadings. Both properties increased with increasing CO₂ loading and decreased with increasing temperature. However, they exhibited different dependencies on solvent composition. Density decreased with increasing AMP concentration but increased with increasing PZ concentration, whereas viscosity increased with the concentration of either AMP or PZ. Empirical correlations were fitted to the collected data and made available for process modeling. Thermodynamic properties, including CO₂ solubility at high temperatures, heat of absorption of CO₂, N₂O solubility, liquid phase speciation, and pH, were also measured. These data provided a broad dataset for developing a thermodynamic framework and understanding the chemistry of the AMP/PZ/H₂O/CO₂ system. The work provided experimental thermodynamic data for CO₂ solubility up to 150 ℃, broadening the temperature and CO₂ loading ranges previously available in the literature. The results show that a small variation in the AMP/PZ ratio has a minimal impact on the thermodynamic properties. However, a larger change in the AMP/PZ ratio results in noticeable variations in the heat of absorption and CO₂ solubility. Liquid speciation data show that CO₂ is prevalently bound as PZ carbamate at low CO₂ loading and as bicarbonate and PZ dicarbamate at high CO₂ loading. These data provide insights into the dominant CO₂ capture mechanism as a function of CO₂ loading. Furthermore, the N2O solubility, which is necessary to estimate the CO₂ physical solubility and for kinetics modeling, was measured and reported as N2O Henry’s constant. The results indicate that the N2O Henry’s constant increases as both the temperature and CO₂ loading increase. The CO₂ absorption kinetics in CESAR1 were measured as functions of temperature and CO₂ loading. The measurements provide a basis for understanding how the absorption kinetics of CESAR1 change with the CO₂ loading. CESAR1 maintained faster absorption kinetics compared to MEA even at relatively high loadings. However, the experimental range did not extend to the highest CO₂ loadings encountered during pilot-plant operation, and further measurements under these conditions and modeling activities are therefore required. The experimental results supported the development of an electrolyte non-random two-liquid (e-NRTL) thermodynamic model for the AMP/PZ/H₂O/CO₂ system implemented in Aspen Plus v14. The thermodynamic framework accurately represents different properties such as CO₂ equilibrium partial pressure, total pressure, liquid speciation, heat of absorption of CO₂, and amine volatility over a broad range of AMP and PZ concentrations, temperature, and CO₂ loading. The thermodynamic framework was then integrated with a mass-transfer and kinetics framework to develop an open-access, rate-based Aspen Plus model for CESAR1. The process model was validated using pilot plant data from the University of Kaiserslautern, the Technology Centre Mongstad, and the Tiller SINTEF pilot plant. The model predicts CO₂ capture, solvent loading, column temperature profiles, and specific reboiler duty with good accuracy over a broad range of operating conditions and plant geometries. The developed model was then used to investigate and consistently compare the performance of MEA and CESAR1 at CO₂ capture rates of up to 99.8% for flue gases representative of different industrial sectors. The results show that CESAR1 could reduce the reboiler duty by up to 20% relative to MEA over the entire range of CO₂ capture rates investigated. The techno-economic analysis indicates that this energy advantage is generally translated into slightly lower capture costs, even though the benefit could be offset by higher capital costs. The validated model was also applied to investigate why MEA and CESAR1 respond differently to selected process modifications. Three modifications were investigated: increased stripper pressure, lean vapor compression, and absorber intercooling. The results demonstrate that the effectiveness of these modifications is solvent dependent. Lean vapor compression benefits MEA more than CESAR1. The reason is that MEA needs to be operated at a higher liquid flow rate than CESAR1 to achieve the same capture rate. This increases the potential latent heat recovery from the lean solution. Furthermore, the energy required to produce the stripping steam is higher for MEA compared to CESAR1. Intercooling efficiency is also solvent dependent. The results indicate that the intercooling benefits MEA more than CESAR1, and that the optimal intercooling location differs between the two solvents. Finally, an increase in reboiler pressure effectively decreases the required regeneration duty, with similar effects for both MEA and CESAR1. Overall, this thesis provides new experimental data, a rigorous thermodynamic framework, and a validated rate-based model for CESAR1. The combined experimental and modeling methodology developed in this work is transferable for evaluating emerging solvents for more efficient amine-based CO₂ capture processes.